Incremental fuzzy clustering with multiple kernels

This paper presents two incremental clustering algorithms based on FCMK, a fuzzy clustering with multiple kernels algorithm we developed earlier [1]. The FCMK algorithm has a memory requirement of O(N²), where N is the number of objects in the data set. Thus, even data sets that have nearly 1, 000, 000 objects require terabytes of working memory-impractical for most computers. One way to attack this problem is by using incremental algorithms; these algorithms sequentially process chunks or samples of the data, combining the results from each chunk. The proposed incremental algorithms neither use any complicated data structure nor any complicated data compression techniques, yet produce data partitions comparable to FCMK. We assess the performance of our incremental algorithms by, first, comparing their clustering results to that of the FCMK and, second, by showing that these algorithms can produce reasonable partitions of large data sets.